HI Observations of Galaxies in the Southern Filament of the Virgo

arXiv:1609.03896v1 [astro-ph.GA] 13 Sep 2016 o.Nt .Ato.Soc. Astron. R. Not. Mon. ivnli&Hye 1985 Haynes & Giovanelli epc otermrhlgclcutrat vligi le in evolving ( counterparts regions morphological their to respect ⋆ stripp RP undergoing cluster Virgo the in galaxies of amples respon (e.g. partly quenching be tion to known is H It the disk. ex- for galaxy gravitationa the sible the the and than from cluster greater is force galaxy ISM ing a the of by felt ICM force the ternal through moves galaxy str (RP) a pressure ram Perh the (e.g. is clusters. ping work in at galaxies mechanism those important of most evolution the processes drive hydrodynamical stru then scale through large (ICM) the medium with tracluster interacts galaxies spiral interste structure of scale (ISM) small the galaxies, of clusters in H be to tend environments dense in evolving ies H .Sorgho A. WSRT the and KAT-7 Pathfinder SKA the with Cluster tudrosdfeetmcaim htdtrieismorph its determine gr that a (e.g. or mechanisms cluster a different as undergoes such region it dense a in resides galaxy a When INTRODUCTION 1 4 3 2 1 c bevtiedAtohsqed ’nvri´ eOuagado l’Université de de d’Astrophysique Observatoire aty srnmclIsiue nvriyo Groningen, of University Institute, Astronomical Kapteyn ehrad nttt o ai srnm ATO) Postb (ASTRON), Astronomy Radio for Institute Netherlands eateto srnm,Uiest fCp on rvt B Private Town, Cape of University Astronomy, of Department [email protected] 00RAS 0000 I rslr1980 Dressler bevtoso aaisi h otenfiaeto h Virg the of filament southern the in galaxies of observations ans ivnli&Cicrn 1984 Chincarini & Giovanelli Haynes, un&Gt 1972 Gott & Gunn I 1 ecec ( deficiency ⋆ n a rpris( properties gas and ) .Hess K. , aoh aar ors2000 Morris & Navarro Balogh, 000 ; 0–0 00)Pitd2 aur 08(NL (MN 2018 January 20 Printed (0000) 000–000 , ole ta.2001 al. et Vollmer ; ohn ulvn&Shme 1982 Schommer & Sullivan Bothun, ole ta.2001 al. et Vollmer 1 , 2 iha H an with empteH the map We ABSTRACT ifrn emszso h w eecps iia column similar a telescopes, two the of sizes beam different ot fteVrocutruigteKAT the using cluster Virgo the of south 0oto h 4glxe rsn H present galaxies 14 the of out 10 e words: Key cluster. not the are in effects elsewhere environmental the that suggesting spirals, 1 S niiul G 4424 NGC individual: – ISM a rsuesrpiga tflstwrstecnr fthe of centre H the 3 towards a and falls is galaxies it 4424 14 as NGC stripping that pressure suggest tail, ram the a of shape the with together ∼ h ag n ml cl tutrs u oa unprecedente an to advantag Out take structures. scale and small observations and the large combine the to approach new a , 3 × aat ta.1990 al. et Cayatte 60 .Carignan C. , 10 p en tipdofNC42,apcla prlglx.Th galaxy. spiral peculiar a 4424, NGC off stripped being kpc 18 tm cm atoms .Mr specifically More ). I n trforma- star and ) asof mass .I cuswhen occurs It ). I ehius nefrmti aais lses–indivi – clusters galaxies: – interferometric techniques: ecetwith deficient I .Tpclex- Typical ). OBx80 70A rnne,TeNetherlands The Groningen, AV 9700 800, Box PO gu(DU) ugduo,BriaFaso Burkina Ouagadougou, (ODAUO), ugou lrmedium llar itiuino aaisi a in galaxies of distribution s2 90A wneo,TeNetherlands The Dwingeloo, AA 7990 2, us gX,Rneoc 71 ot Africa South 7701, Rondebosch X3, ag .Galax- ). tr in- cture sdense ss − I restor- l 7 p the aps 2 luslcigotclcutrat.Oeo h lusi n a is clouds the of One counterparts. optical lacking clouds × ology 1 a ece ihtetoosrain vr1. ms km 16.5 over observations two the with reached was oup, , that 10 ing 4 ip- .A Oosterloo A. T. , - ; 7 M ⊙ n togH strong a and ac fM87, of tance c of radius virial the to out m galaxies ( important of stripping an gas remains of it nism clu that of reveal regions central simulations the mological in strongest is effect this though ( 4402 ( ntermrhlg n ailvlcte,a remembers true as 1987 velocities, Sandage (e.g. & Tammann Cluster radial Binggeli, Virgo and the morphology of their on with and IGM the with all interactions neighbours. their environment, probe surrounding directly their to us and th map galaxies to of opportunity an regions provides gas neutral the particular nld G 38( 4388 NGC include tansfrmdligglx vlto (e.g. evolution galaxy simulations modelling numerical for for straints tests provide i quenching they formation Further, star ies. as well as growt disks, the gas of of understanding depletion our atomi to p key neutral complete are the environments a of dense from Observations evolution. far galaxy still of are ture we understood, well seem ments onsn ra a okm2007 Gorkom van & Bryan Tonnesen, 2004 Vollmer & Gorkom van Kenney, I odt,aot10 aaishv enietfid based identified, been have galaxies 1300 about date, To hl h pia hrceitc fglxe ndneenvi dense in galaxies of characteristics optical the While A ecece o ihrta hs ftecutrslate cluster’s the of those than higher not deficiencies rw ta.2005 al. et Crowl T E tl l v2.2) file style X − n h STitreoees eas fthe of Because interferometers. WSRT the and 7 ∼ el ta.2014 al. et Neill 1 I . rfiewith profile 5 2 ◦ , 3 × otro a okm2005 Gorkom van & Oosterloo n G 48( 4438 NGC and ) 2 . 5 oepoone ntergo than region the in pronounced more ◦ adg,Bngl amn 1985 Tammann & Binggeli Sandage, ig lse.W eetattlof total a detect We Cluster. Virgo einlctda h iilradius virial the at located region xet edtc nH an detect we extent, d ,tercsinvlct fteclus- the of velocity recession the ), W otmre aayundergoing galaxy post-merger est estvt of sensitivity density 50 ftersniiiyt both to sensitivity their of e 73 = .Lctda 68Mc(dis- Mpc 16.8 at Located ). rpriso h galaxy, the of properties e ; ). ole ta.2004 al. et Vollmer ul ig galaxies: – Virgo dual: hmne l 2005 al. et Chemin ms km a´ ta.2013 al. Davé et − − 1 1 efidthat find We . epioneer We . wdetection ew ,NC4522 NGC ), o tr,cos- sters, n con- and I N galax- n a in gas c ,NGC ), outer e alof tail lusters owing H .Al- ). echa- and h their .In ). I ron- ic- ∼ ; 2 Sorgho et al. ter is 1050 km s−1, and its velocity dispersion is 700 km s−1∼(Binggeli, Popescu & Tammann 1993). Detailed X-ray∼ observations of the hot cluster gas (Böhringer et al. 1994) revealed several substructures, suggesting that the cluster is not virialised, as seen in other clusters (see e.g. for the Coma & Antlia clusters, Briel, Henry & Boehringer 1992; Hess et al. 2015). The Virgo Cluster, instead, is still growing. Subclusters around the ellipticals M86 and M49 (whose systemic velocities are respectively -181 and 950 km s−1, Kim et al. 2014) are merging with the main cluster around the central elliptical M87. Several HI imaging surveys of the cluster revealed that the HI disks of its central late-type galaxies are truncated to well within the optical disks, suggesting that ICM - ISM interactions play an important role in driving the evolution of galaxies in the inner region of the cluster (Warmels 1988; Cayatte et al. 1990; Chung et al. 2009; Boselli et al. 2014). Wide-area HI surveys like HIPASS (HI Parkes All-Sky Survey, Barnes et al. 2001), ALFALFA (Arecibo Legacy Fast ALFA, Giovanelli et al. 2005) and AGES (Arecibo Galaxy Environment Survey, Auld et al. 2006) done with single- dish telescopes have helped detect many low HI mass galaxies such as early-type and late-type dwarf galaxies, but also starless “dark” galaxies and several HI clouds (e.g., Alighieri et al. 2007; Kent et al. 2007). In particular, ALFALFA observations allowed the detection of complex structures associated with Virgo galaxies (e.g. Koopmann & Giovanelli 2008), showing probable signs of interactions in the cluster. We use the KAT-7 and the WSRT to study galaxies located in an X-ray filament to the South-West of the cluster connecting the main cluster with the substructure associated with M49. We Figure 1. X-ray map of the Virgo Cluster (0.5 − 2.0 keV, ROSAT; observe a 1.5◦ 2.5◦ region in the M49 subcluster. It is located Böhringer et al. 1994). The circle shows the virial radius of the cluster ∼ × ◦ . ◦ south of the Virgo Cluster, 3 away from the elliptical M87, (∼ 3 7 or 1.08 Mpc, Arnaud, Pointecouteau & Pratt 2005; Urban et al. 2011) around M87, and the box is the region observed in the present work. and contains the elliptical M49∼ (see Fig. 1). The field also contains The labelled plus signs show the major ellipticals of the cluster, and the the one-sided HI tail galaxy, NGC 4424, previously observed using cross indicates the position of the HI-tailed galaxy NGC 4424. the VLA (Chung et al. 2007, 2009). Taking advantage of the short baselines of the KAT-7 telescope sensitive to extended structures and the high spatial resolution of the WSRT telescope, we map late- to calibrate the phase stability. The choice of these calibrators lies type galaxies in the region down to low column densities. Given the on their close angular distance from the target field and their good large difference in the resolutions of the two arrays, we develop a quality status as given by the ATCA Calibrator Database1 and the new technique to combine the data from the two telescopes. VLA Calibrator Manual2. The calibrator must be as close as possi- We begin with a description of the observations with the two ble to the target, and have a good quality rating in the 21cm L-band. telescopes, as well as the technique adopted for the data combina- We used three pointings to cover the desired field, uniformly tion in section 2. The results are presented in section 3 where the separated (by 72% of the primary beam) and following the line HI map of the region as well as the HI parameters of the detections connecting the∼ objects NGC4424 and M49 (see Fig. 2). The upper are given. In section 4 we discuss the results in light of the cluster pointing was observed for 30 hours split in 6 sessions, while the environment effects, and give a summary in section 5. The contours two others were given 24∼ hours each during a total of 8 sessions. and HI profiles of all the detections are presented in the Appendix. Initially, the observations were focused on NGC 4424 and its HI tail; however, to have a complete view of the tail and to probe the whole region between the galaxy and the elliptical M49, the upper pointing was expanded to a mosaic. 2 OBSERVATIONS AND DATA REDUCTION During these observations, KAT-7 was still in the commis- 2.1 KAT-7 observations and reduction sioning phase and the RFI monitoring was still ongoing, therefore special care needed to be given to the inspection of the data The Karoo Array Telescope (KAT-7) observations were conducted beforehand. Manual flagging was performed to remove RFI, and between July 2013 and May 2014 during a total of 14 sessions. special care was taken to suppress the effects of Galactic HI on All observations were centred at a frequency of 1418 MHz and the sources of interest. The standard CASA package (version 4.1.0, correlator mode provided a total bandwidth of 25 MHz, covering McMullin et al. 2007) was used to image the data. The observ- − the velocity range 2115 to 3179 km s 1 divided in 4096 channels ing sessions were individually reduced, and continuum subtrac- − −1 being 6.1 kHz (1.28 km s ) wide. tion was performed by using a first order fit over line free chan- The observation technique used with KAT-7 consists in alter- nating between the target and a phase calibrator (1254+116). The flux calibrator (either PKS 1934-638 or 3C 286) was observed two 1 http://www.narrabri.atnf.csiro.au/calibrators/ or three times per session and the phase calibrator every 20 min 2 https://science.nrao.edu/facilities/vla/docs/manuals/cal

c 0000 RAS, MNRAS 000, 000–000 Observations of Virgo Southern Filament 3

are observed only at the beginning and at the end of the individual 12h observing runs to fix the flux scale. To be more sensitive to extended features, the Maxi-Short configuration was adopted for the observations, and the centres of the individual pointings ( 22′ apart) were chosen, like for the KAT-7 observations, such th∼at the sensitivity is uniform across the field (see Fig. 2). The L-band correlator used provides a 20 MHz band split over 1024 channels, i.e, 4.12 km s−1 per channel. We reduced the WSRT data using MIRIAD (Sault, Teuben & Wright 1995), and following the standard procedure for WSRT data reduction. Gain calibration was achieved by self-calibrating on background continuum sources throughout the observed field. Continuum subtraction was performed with the uvlin task in the u,v plane and the solutions of the self-calibrated continuum data were subsequently applied to the calibrated line data. Next, the spectral regions of interest were selected and every 4 channels were averaged, changing the spectral resolution from 4.12 km s−1 to 16.49 km s−1. The individual observations were then co-added into a single u,v dataset. A mosaic datacube was produced using the Robust weighting with a robustness R = 0.5, and a sampling interval (pixel size) of 5′′. A Gaussian taper of 30′′ was also applied. Because of the East- West layout of the WSRT array and the low declination of the field, the restoring beam is very elongated: 2.9′ 0.6′. The datacube × was then cleaned in MIRIAD using MOSSDI, the cleaning task for mosaics. The cleaning was performed only in channels containing emission, and in regions around emission features. A primary beam Figure 2. The KAT-7 (green) and WSRT (purple) primary beams on an correction was finally applied to the cube. The third column of Ta- optical SDSS r-band grayscale, showing the pointing positions of the observations. The field is the same as the box in Figure 1. ble 1 summarises the parameters of the WSRT image cube.

2.3 Combination of the two datasets Parameter KAT-7 WSRT Data combination is widely used in radio astronomy, to get Weighting function Robust = 0 Robust = 0.5 the most out of observations. Several methods have been devel- Channel width 15.2 km s−1 16.5 km s−1 oped, performing the combination in either the image domain FWHM of beam 4.4′ 3.1′ 2.9′ 0.6′ or in the u,v plane (see e.g., Stanimiroviˇc2002). Techniques of × ◦ × ◦ PA of beam 8 1 combining data from different telescopes are generally used for ∼−′′ ′′ ′′∼ ′′ Map gridding 55 55 5 5 short spacing correction, i.e, used to combine single-dish data ◦ × ◦ ◦ × ◦ Map size 3.9 3.9 1.7 2.3 to interferometer data (e.g., Vogel et al. 1984; Schwarz & Wakker × −1 × −1 RMS noise 2.5 mJy beam 0.35 mJy beam 1991; Stanimiroviˇcet al. 1999; Kurono, Morita & Kamazaki 2009; −1 18 −2 18 −2 NHI (1σ & 16.5 km s ) 0.9 10 cm 1.0 10 cm Faridani et al. 2014). As for the combination of two interferome- × × ter data from different arrays, the procedure is quite different (see e.g. Carignan & Purton 1998). The straightforward way of combin- Table 1: Summary of the KAT-7 and WSRT image cubes. ing interferometer data from different arrays is adding them in the u,v plane. This works best when the individual data have the same sensitivity, which is not always the case. nels. Because KAT-7 does not use Doppler tracking and CASA While the average rms noise in the KAT-7 data is 4.1.0 does not fully recognise frequency keywords, special care was −1, that in the WSRT is much smaller, ∼ taken to produce image cubes with the proper velocity coordinates 2.5 mJy beam −1. Since the weight assigned to the datasets dur-∼ (see Carignan et al. 2013). The parameters of the resulting primary 0.35 mJy beam ing combination goes as 2 , the WSRT data end up having a beam-corrected data cube are summarised in Table 1. 1/σrms much larger weight relative to the KAT-7 data. If IK , wK are respectively the intensity and weight of a given pixel in the KAT-7 2.2 WSRT observations and reduction data and IW , wW those of the corresponding pixel in the WSRT data, the intensity of the corresponding pixel in the resulting com- A total of 4 sessions of 12h observation each were conducted bined data is in April 2006 with the Westerbork Synthesis Radio Telescope IK wK + IW wW (WSRT) single-feed system, prior to the installation of APERTIF Ic = . (1) wK + wW (Aperture Tile In Focus). An individual pointing was observed dur- 2 ing each session; two calibrators were also observed, 3C 147 and So, for w 1/σ , wK wW and Ic IW . ∝ rms ≪ ∼ CTD 93. Unlike the KAT-7 observations, WSRT observations do Therefore, the traditional combination in the u, v plane using not require gain calibrators at regular time intervals. This is because the dedicated MIRIAD tasks (see e.g Hess et al. 2009) results in a the phase of the WSRT system is quite stable, and the calibrators largely WSRT feature-dominated data, with the KAT-7 features be-

c 0000 RAS, MNRAS 000, 000–000 4 Sorgho et al. ing essentially nonexistent. To avoid this, we performed the combination in the image plane based on column density sensitivity. This procedure provides a complex ﬁnal beam that has a double gaussian shape. The ratio between the rms noises is more or less equal to the ratio between the respective synthesised beam areas. The column density of a source is given by

i,j 21 IHI NHI /∆v = 1.1 10 (2) × bmin bmaj or, in terms of brightness temperature,

i,j IHI TB = 611 , (3) bmin bmaj where bmin and bmaj are respectively the minor and major FWHM i,j of the synthesised beam in arcsec, and IHI the HI intensity of a pixel of coordinates (i,j) in mJy beam−1. The brightness temperature TB is expressed in K. The column density is proportional to the flux and inversely proportional to the beam area. Therefore, the two datasets end up having the same order of column density sensitivity. The most appropriate way of combining them is by converting the cubes into units of column density per unit velocity in the image plane. The continuum-subtracted KAT-7 u,v data was first exported from CASA to MIRIAD via UVFITS format. Once imported into MIRIAD, the velocity axis was resampled to match that of the WSRT data in velocity range and width. This was achieved with the UVAVER task. The spatial centre of the mosaic was set to be that of the WSRT cube, and a gaussian taper of 30′′ was also applied. The resulting cube was cleaned and primary beam corrected using the MOSSDI and LINMOS tasks respectively. However, since Figure 3. Map showing the variation of the noise in the combined cube. the KAT-7 field of view is larger than that of the WSRT, we used The five green dots show the positions of the HI clouds detections reported REGRID to regrid the spatial axes of the WSRT cube to the KAT-7 in Kent et al. (2007) and discussed in section 3.3. new cube. At this stage, both image cubes have the same dimensions, co- respectively along the major and minor axes. Assuming that posi- ordinate grids, reference pixel and pixel size. This way, any pixel in tion angle of the two beams are equal (Table 1 shows that they are one cube has the same world coordinates as its corresponding pixel similar in the present case), we can sum the two gaussian functions in the other cube. Equation 2 was then used, with the MATHS task, and the major and minor axes of the resulting function can respec- to convert them into column density cubes. Since the cubes initially 2 2 0.5 tively be written ∆αc = [(0.56 ∆αK ) + (0.44 ∆αW ) ] and had units of flux and Equation 2 uses the line flux, the output cubes 2 2 0.5 − − − ∆δc = [(0.56 ∆δK ) +(0.44 ∆δW ) ] . Using the parameters of have units of column density per unit velocity cm 2 (kms 1) 1 the individual beams in Table 1, we obtain the approximate param- or, using Equation 3, units of brightness temperature K. The aver- eters of the combined beam to be 2.8′ 1.7′. age rms noise levels (measured in the central regions of the cubes × and over line-free channels) are 31 mK and 5 mK respectively in the output KAT-7 and WSRT data cubes: a weight ratio of 1.26. Finally, the MATHS task was used to linearly combine the two 3 RESULTS data cubes, with the expression 3.1 HI maps and HI properties

1.26 IK + IW Ic = . (4) Figure 4 shows the HI total intensity maps of all detections made 2.26 in the field, overlaid on an X-ray map (0.5 2.0 keV, ROSAT; 17 −2 − The rms noise in this final cube is 27 mK (or 7.9 10 cm ) at a Böhringer et al. 1994) of the cluster. A total of 17 objects were de- velocity resolution of 16.5 km s−1 and allowed the× detection of 12 tected, including 14 galaxies with confirmed optical counterparts objects at the 3σ level. In Figure 3 we present a map of the noise and 3 HI clouds. The HI detection of 2 of the 3 HI clouds was in the combined data, showing the variation of the noise across the previously reported, but they have not been optically detected as field. To be sensitive to low surface brightness objects, the data they have no counterparts in the SDSS. As for the third cloud, − cube was furthermore Hanning-smoothed to 33 km s 1 with an no HI nor optical detection has been reported in the literature. rms noise of 23 mK (7.0 1017 cm−2), and 5 more objects were We refer to this as the KW (KAT-7+WSRT) cloud. We discuss detected in this new cube.× these individual clouds in section 3.3. In Table 2 we present the To obtain the approximate size of the final beam in the com- HI properties of these different detections as determined from the bined data, we consider two individual two-dimensional gaussian combined image cube. The columns (1), (2) & (3) contain re- functions of standard deviation σα and σδ each and along each axis, spectively, the common names of their optical counterparts when representing the beams of the initial datasets and whose FWHMs available, their optical J2000 coordinates and morphological types can each be written ∆α = 2√2 ln 2 σα and ∆δ = 2√2 ln 2 σδ, when available. In columns (4) and (5) we list respectively the

c 0000 RAS, MNRAS 000, 000–000 Observations of Virgo Southern Filament 5 optical diameters at a surface brightness of 25 mag arcsec−2 in the B band, and inclinations as obtained from the RC3 catalogue (Corwin, Buta & de Vaucouleurs 1994). In column (6) we list the Tully-Fisher distances used to estimate the HI masses, mostly from Solanes et al. (2002). In columns (7) and (8) we list, respectively, the systemic velocities and velocity widths obtained from the combined image cube. In columns (9) and (10) we give the measured HI masses and deﬁciency parameters of the detections along with error estimates, and ﬁnally, in column (11) their projected angular distances from M87.

3.2 NGC 4424 HI tail The main goal of combining the two data sets was to achieve better HI column density sensitivity and detect fainter structures. In the case of the galaxy NGC 4424, we seek to map the full extent of the HI tail. In Fig. 5 we present a comparison of the HI map of the galaxy as obtained in the three different datasets (KAT-7, WSRT & combined). While the tail is seen in the KAT-7 data, its southern end is not detected in the WSRT data. When combined, the data reveal the structure of the tail with a better sensitivity of NHI 17 −2 −1 ∼ 8 10 cm over 16.5 km s . Assuming a distance of 15.2 × Mpc (Cort´es, Kenney & Hardy 2008), the tail extends out to 60 ∼ kpc from the main envelope, with an HI mass of (4.3 0.9) 7 ± × 10 M⊙. This represents 20% of the total HI mass of the galaxy. In previous VLA observations∼ of the galaxy by Chung et al. (2007) down to 2 1019 cm−2, only 20 kpc of tail was seen. The ∼ × ∼ present observations therefore reveal an HI tail three times longer than previously observed.

3.3 Gas clouds

Three HI clouds were detected in the combined image cube: the M49 cloud spatially located near M49 and at a systemic velocity −1 −1 of 476 km s , the KW cloud located at vsys = 1270 km s and whose spatial coordinates coincide with NGC 4451 but with an offset of 400 km s−1 in velocity (see Figure 4), and Cloud 7c (aka AGC 226161), previously observed by Kent et al. (2007). The M49 cloud lies, both in velocity and space, between the late-type dwarf galaxy UGC 7636 (at 276 km s−1) and the giant − elliptical M49 (at 950 km s 1). Previous HI and optical observations of these galaxies revealed that the cloud is a result of the combined effect of gas stripping due to RP and tidal interaction between the dwarf galaxy and the elliptical (Henning, Sancisi & McNamara 1993; McNamara et al. 1994). The gas cloud would have been stripped from the ISM of the dwarf galaxy due to M49. Later multi-wavelength observations by Arrigoni Battaia et al. (2012) confirmed these results. It was also detected in the AGES Survey (Taylor et al. 2012). As for the KW cloud, no previous detection was reported in the literature, and follow-up observations are needed to confirm the lack of a stellar counterpart. The cloud has an HI mass of (7 1) 7 ± × 10 M⊙ at the distance of M87, and presents a Gaussian profile of −1 width W50 = 73.2 km s . Figure 5. NGC 4424 as seen in the KAT-7 (top), WSRT (middle) and 7 The cloud 7c, detected with MHI = (6 1) 10 M⊙ (as- combined (bottom) data, overlaid on SDSS r-band grayscale. The con- ± × −1 suming a distance of M87) and a velocity width of 74 km s , is tour levels of the KAT-7 and WSRT maps are respectively 0.7 (3σ), 1.0, one of the 8 objects that Kent et al. (2007) reported to not have 2.0, 3.0, 4.0, 5.0 × 1019 atoms cm−2 and 0.75, 1.5, 3.0, 6.0, 12.0, stellar counterparts, and qualified as “optically unseen detections”. 15.0 × 1019 atoms cm−2, and those of the combined map are 0.5, 1, 2, 3 19 −2 The authors, who reported an HI mass of the cloud almost 3 times ... 10 ×10 atoms cm . higher than what is measured here, argue that 7c is part of a complex of 5 HI clouds – cloud 7 – among which only 2 (7c and 7d)

c 0000 RAS, MNRAS 000, 000–000 6 Sorgho et al.

Figure 4. The zoomed-in observed region (box of Fig. 1). The black contours are the X-ray levels (4.2, 4.4, 4.7, 5.0, 5.2, 5.5 & 5.8 × 10−11 ergs cm−2 s−1). The coloured contours are HI column densities (as obtained from the combined map), ranging from 5 × 1018 to 8 × 1020 cm−2 (see Figures A1-A13 for −1 individual contours levels) . For reference, the giant elliptical M49 is at vsys = 950 km s . The dotted line shows the virial radius of the cluster.

7 −1 are confirmed by follow-up VLA observations (Kent 2010). The 3), the detection limit goes up to 1.1 10 M⊙/km s . The AL- × authors, however, explain that the remaining 3 objects were not FALFA HI masses reported in Kent et al. (2007) for the five clouds expected to be confirmed by the follow-up observations because put 7a, 7d and 7e below the detection limits reported above, even either distant from the pointing centre and therefore below the 3σ if one adopts a velocity width as low as 50 km s−1. However, the detection limit (7b & 7e) or outside the field of view (7a). At the 7b cloud is well above the detection limit and it comes as a sur- positions of the clouds 7b and 7c, the noise level in the present ob- prise that the cloud was not detected. This fact, along with the low servations is about 89 mK (or 1.6 1017 cm−2(kms−1)−1), which detected mass of 7c, suggests that some low column density gas is × 6 −1 gives a 3σ detection limit of 1.25 10 M⊙/km s . The clouds being missed in the region. 7d and 7e are located at the edge of× the map, and the detection limit Although not contained in any major catalogue, a matching 6 −1 at their position increases to 3.2 10 M⊙/km s . At the position optical redshift for AGESVC1 293 was previously associated with × of 7a which coincides with the edge of the WSRT beam (see Figure an optical counterpart which has an optical redshift in the SDSS

c 0000 RAS, MNRAS 000, 000–000 Observations of Virgo Southern Filament 7 database (Taylor et al. 2012). We therefore do not classify the ob- as (Solanes et al. 2002, hereafter Sol02) ject as an HI cloud, but do not include it in the analysis in section defHI = log ΣHI (D,T ) log ΣHI , (5) 4.2 since its optical properties are not clearly determined. h i− where the mean hybrid HI surface density is obtained from

the ratio of the integrated HI ﬂux density SHI of the galaxy −1 4 DISCUSSION in Jy km s and its apparent optical diameter a in arcminutes (Solanes, Giovanelli & Haynes 1996):

4.1 HI morphologies 2 ΣHI = SHI /a . (6) The Virgo Cluster is not dynamically re- The term log ΣHI (T ) is the base-10 logarithm of the expected laxed (Binggeli, Tammann & Sandage 1987; h i Binggeli, Popescu & Tammann 1993), but contains different hybrid HI column density value for a field galaxy of the same di- substructures with the largest being the subclusters A (centred on ameter D and morphological type T . The higher the deficiency pa- M87) and B (centred on M49). The ensemble of late type galaxies rameter, the more HI-poor the galaxy is relative to the mean of a in the cluster presents a larger velocity dispersion than early types, comparison sample of field galaxies. Column (10) of Table 2 lists suggesting that many of the late types are in the stage of infall the values of defHI for the detected galaxies. Among the 13 de- (Tully & Shaya 1984; Binggeli, Tammann & Sandage 1987). tected galaxies with clear morphological type and having optical It is well known that the infall of the galaxies in a cluster data, only 3 (namely UGC 7590, VCC 1142 & 0952) are not HI- is likely to happen along filaments (e.g., Porter & Raychaudhury deficient. 2007; Mahajan, Raychaudhury & Pimbblet 2012). Furthermore, To put these quantities into context, we compare the HI de- the distribution of the hot X-ray in the Virgo cluster (see Fig. 1) ficiency of the detections to the VLA Imaging survey of Virgo suggests a filamentary structure connecting the subclusters A and galaxies in Atomic gas (VIVA, Chung et al. 2009) sample, and to B, and the distribution of the detected galaxies in the region seems the sample of “Local Orphan Galaxies” (LOG, Karachentsev et al. to follow this filament. In Fig. 4 we present the observed region 2011). The VIVA sample consists of 53 late type galaxies, mostly selected to cover a range of star formation properties based on of Fig. 1, showing an overlay of the HI maps of detections on the hot X-ray gas in the region. Most of the detections belonging (spa- Koopmann & Kenney (2004)’s classification, and both in high and low density environments. The galaxies span the angular distances tially) to the filament present an asymmetry to the south in their HI of 1◦ to 12◦ from M87 (out to 1.6 Abell radii of the clus- morphologies; these are NGC 4424, NGC 4451, NGC 4445, UGC ∼ 7596 and AGESVC1 293. All these galaxies have systemic veloc- ter). Chung et al. (2009) provide the HI masses and distance- ities comprised between 400 and 900 km s−1, and are blueshifted independent deficiency parameters for all the galaxies except for with respect to the systemic velocity of the cluster ( 1050 km two (IC 3418 and VCC 2062) where the HI deficiency is not de- s−1). They seem to be falling inside the cluster from the∼ back, along rived. The LOG catalog is made of 517 local isolated galaxies se- the filament. 3 4 Previous works showed that RP causes cluster galaxies to be- lected from HyperLEDA and NED . Of these, 278 are classified as irregular galaxies, known to be gas-rich. Of the remaining 239 gin forming HI tails at about the virial radius (e.g., Chung et al. 2007). Fig. 4 shows that most of the detections that are projected in spirals, only 186 have their optical sizes listed in HyperLEDA. We the filament are located about a virial radius from M87, and might hereafter refer to these as the LOG sample for our comparisons. well be subject to RP stripping. However, except for NGC 4424, The optical B-band diameters (at the 25th magnitude) of the LOG galaxies, combined with their HI line flux, distance and morpho- none of them exhibit obvious HI tails. If the asymmetry observed in the galaxies is a result of RP logical type provided in Karachentsev et al. (2011), were used to stripping due to the relative movement of the galaxies through the evaluate their HI deficiency parameters. The HI masses were de- hot halo of the main cluster of Virgo, one should expect a corre- rived using lation between the asymmetry and the distance to M87. However, 5 2 MHI = 2.36 10 d SHI (7) RP stripping from the M87 halo may not be the sole explanation × for the range of HI morphologies, but some of the galaxies may where d is the distance of the galaxy in Mpc and SHI is the flux in have gone through the M49 halo, causing their HI morphology to Equation 6. be disturbed. In Fig. 6 we present the HI deficiency parameter as a func- The two-dimensional distribution of the detections suggests tion of HI mass for our detection sample over-plotted on the VIVA that they are mostly located in the filamentary structure defined by and LOG samples. The HI deficiency presents little variation with the X-ray distribution in Fig. 4. Their systemic velocities listed in the HI mass for the field galaxies (the LOG sample); it slightly Table 2 also suggests that all the detections, except NGC 4470, be- decreases as one moves to higher HI masses, but remains close to long to the Virgo cluster. However, considering their Tully-Fisher defHI = 0 throughout the probed HI mass range. However, for the distances listed in column (6) of Table 2, it is clear that NGC 4451, galaxies in the vicinity of the Virgo Cluster (the VIVA and the de- NGC 4411B, NGC 4416, NGC 4466 and VCC 0952 are not in the tection samples), the deficiency parameter decreases with increas- filament. They appear to be background galaxies, although their ing HI mass. systemic velocities may suggest otherwise. Apart from these galax- Out of the three non-deficient galaxies, only UGC 7590 has ies, all the other detections may well belong to the filament. a high HI mass. The other two detections, VCC 1142 and VCC 8 0952, have HI masses . 10 M⊙. Figure 4 shows that these galaxies are not confined in a same region, VCC 1142 being near the 4.2 Gas content

To evaluate the relative gas content of the detections, we determine 3 http://leda.univ-lyon1.fr their distance-independent HI deﬁciency parameter; this is deﬁned 4 http://ned.ipac.caltech.edu

c 0000 RAS, MNRAS 000, 000–000 8 Sorgho et al. virial radius and VCC 0952 well within. However, their Tully- Fisher distances distribution (see Table 2) suggest that VCC 0952 is a background galaxy, making it normal for the galaxy to be gas- rich. The relatively high HI mass of UGC 7590 may explain the non-deficiency of the galaxy, as it falls in the lower-right corner of the plot in Figure 6. The abundance of HI in VCC 1142 is a sur- prise, given the location of the galaxy in the cluster (Figure 4) and its low HI mass. Its non-deficiency suggests that, unlike most of the detections, the galaxy is experiencing no HI depletion, but may be accreting gas from the IGM. The Virgo galaxies at the low-mass end of the sample present much higher deficiencies, with respect to the field galaxies, than HI galaxies at the high-mass end. If the samples are representative of the galaxies inside the Virgo cluster, this trend may be an indication def that most of the cluster’s low HI mass galaxies are HI-deficient, suggesting that the cluster environment of Virgo is more severe on low HI mass galaxies. Most of the galaxies in our detection sample present deficiencies lower than the average of the VIVA sample. If they were actual members of the filament connecting the M87 group to the M49 group, one would expect a depletion of their HI content due to interactions with the hot X-ray gas, i.e, they should present higher HI deficiencies with respect to the other cluster members. 7 In fact, the galaxies in the VIVA sample were selected to rep- log MHI (M ⊙) resent late type Virgo spirals in morphological type, systemic velocity, subcluster membership, HI mass, and deficiency. Given that Figure 6. HI deficiency parameter as a function of the HI mass for the most of our detections (including those believed to belong to the different samples. The lines represent the averaged bins. The cyan dots and filament) lie below the average VIVA galaxies in Fig. 6, we con- blue line are the LOG sample, and the orange dots and the red line show clude that the HI depletion of the galaxies in the filament is not the VIVA sample. The black symbols represent the detections in this work more pronounced than those in the M87 halo. (the crosses are the 5 background galaxies discussed in section 4.1, and the diamonds are believed to belong to the filament), and the purple star denotes the one-sided HI-tail’ed spiral NGC 4424. 4.3 Case of NGC 4424 Several authors argued that the tail seen in NGC 4424 is a mixed effect of RP and tidal interaction (Kenney et al. 1996; Cortés, Kenney & Hardy 2006; Chung et al. 2007, 2009). The ar- Although the authors identified the galaxy NGC 4192 as a possible guments for tidal interaction are mainly based on the optical mor- interacting companion for VIRGOHI 21, they also point that the phology of the galaxy, classified as a chaotic early spiral in the high velocity of the galaxy could move it relatively far for the HI Uppsala General Catalogue (UGC, Nilson 1973), a peculiar Sa by tail. This makes it hard to identify the real interactor, and any mas- Binggeli, Sandage & Tammann (1985), and an uncertain barred Sa sive galaxy in a certain radius could be a potential candidate. This by de Vaucouleurs et al. (1991). The Hα morphology of the galaxy type of process might explain the single tail of NGC 4424, and the has been described as truncated/compact by Koopmann & Kenney galaxy could have collided with any massive galaxy in the Virgo (2004) and optical observations reveal that it has a heavily disturbed cluster. stellar disk, and shows a strong Hα emission confined to its central Chung et al. (2007) compared the RP to the restoring force kpc (Kenney et al. 1996; Cortés, Kenney & Hardy 2006). in HI-tail’ed Virgo galaxies using the simulations of Vollmer et al. However, the present observations reveal an HI tail of NGC (2001) and found that for NGC 4424, RP due to the observed hot 4424 three times longer than previously thought. This fact ques- X-ray emitting gas could exceed the restoring force, meaning that tions the hypothesis of tidal interaction as the main cause of the RP is likely at the origin of the HI tail. tail, given that the optical morphology does not leave room for Several observations and simulations have shown that RP may such a long tail since there is no stellar counterpart to the tail, such be effective in removing gas from cluster galaxies out to 1 2 as one would expect from a tidal encounter. Moreover, tidal in- virial radii (e.g. Kenney, van Gorkom & Vollmer 2004; Crowl et− al. teractions usually create two tails (e.g., Toomre & Toomre 1972; 2005; Tonnesen, Bryan & van Gorkom 2007; Bahéet al. 2013). In Yun, Ho & Lo 1994), while we only detect one long tail here. Fig. 4 we show that NGC 4424 is well within the virial radius of the The “dark” galaxy VIRGOHI 21 was observed in the Virgo cluster ( 0.8 rvir) and therefore may be subject to RP stripping. ∼ cluster to have a similar HI tail; it’s a starless HI cloud, appar- By comparison, the only evidence for tidal interaction in NGC 4424 ently connected by a faint HI bridge to NGC 4254, a spiral galaxy is its complex optical morphology. We argue that this is not suffi- located 20′ away (Minchin et al. 2007). Later numerical sim- cient to explain the origin of the tail. A possible scenario would be ∼ ulations by Duc & Bournaud (2008) showed that the long tail of that a tidal interaction moved the galaxy’s HI to a larger radius, the cloud might have resulted from a high-velocity collision, prob- making it easier for RP to ‘blow away’ the gas. This would make ably with the parent galaxy NGC 4254. The simulations predict the RP stripping more efficient, and could well explain the high 7 the formation of a short counter-tidal tail that falls back onto the mass of the HI tail of (4.3 0.9) 10 M⊙, i.e 20% of the ± × ∼ parent galaxy, giving a single-tailed morphology to the structure. galaxy’s total HI mass.

c 0000 RAS, MNRAS 000, 000–000 Observations of Virgo Southern Filament 9

5 SUMMARY AND CONCLUSIONS Binggeli B., Popescu C. C., Tammann G. A., 1993, A&AS, 98, 275 In this work we used the KAT-7 and WSRT telescopes to observe Binggeli B., Sandage A., Tammann G. A., 1985, AJ, 90, 1681 a region of 2.5◦ 1.5◦ located in the Virgo Cluster, 3◦ away Binggeli B., Tammann G. A., Sandage A., 1987, AJ, 94, 251 from the centre∼ of× the cluster and containing the elliptical∼ M49. Böhringer H., Briel U. G., Schwarz R. A., Voges W., Hartner G., The distribution of the hot X-ray gas in the field suggests a fila- Trümper J., 1994, Nature, 368, 828 mentary structure joining the M49 subcluster to the Virgo cluster. Boselli A. et al., 2014, A&A, 570, A69 With a total of 78 and 48 hours of observations respectively with the KAT-7 and∼ WSRT telescopes, we reached similar column Bothun G. D., Sullivan W. T. I., Schommer R. A., 1982, AJ, 87, densities sensitivities with the two telescopes. To detect both the 725 low and high resolution features and benefit from the advantages Briel U. G., Henry J. P., Boehringer H., 1992, A&A, 259 of both telescopes, we combined the two observations. However, Carignan C. et al., 2013, AJ, 146, 48 because of the large difference in the point source flux sensitivi- Carignan C., Purton C., 1998, ApJ, 506, 125 ties in the two datasets, the combination could not be done in the Cayatte V., Balkowski C., van Gorkom J. H., Kotanyi C., 1990, Fourier plane, as is widely practised. We have pioneered a new ap- AJ, 100, 604 proach, which consists of combining the data in the image plane Chemin L. et al., 2005, A&A, 436, 469 after converting the cubes from units of flux density to units of col- Chung A., van Gorkom J. H., Kenney J. D. P., Crowl H., Vollmer umn density. This provided us a dataset highlighting features of B., 2009, AJ, 138, 1741 both the large and small scale structures, and in which we reach an Chung A., van Gorkom J. H., Kenney J. D. P., Vollmer B., 2007, improved 1σ sensitivity of 8 1017 cm−2 over 16.5 km s−1. ApJ, 659, L115 ∼ × A total of 14 galaxies and 3 HI clouds with HI masses Cortés J. R., Kenney J. D. P., Hardy E., 2006, AJ, 131, 747 7 MHI > 10 M⊙ were detected, including the one-sided HI-tailed Cortés J. R., Kenney J. D. P., Hardy E., 2008, ApJ, 683, 78 spiral NGC 4424. 10 out of the 14 galaxies were found to be HI- Corwin H. G. J., Buta R. J., de Vaucouleurs G., 1994, AJ, 108, deficient. Compared to both a sample of other Virgo spirals and 2128 field galaxies, most of the detected galaxies are found to be no more Crowl H. H., Kenney J. D. P., van Gorkom J. H., Vollmer B., 2005, HI-deficient than typical Virgo galaxies, suggesting that gas strip- AJ, 130, 65 ping processes in the region are not more pronounced than else- DavéR., Katz N., Oppenheimer B. D., Kollmeier J. A., Weinberg where in the cluster. In particular, NGC 4424 has an HI-deficiency D. H., 2013, MNRAS, 434, 2645 typical of the cluster galaxies and exhibits a tail observed to be 60 de Vaucouleurs G., de Vaucouleurs A., Corwin, H. G. J., Buta kpc, three times longer than previous observations with the VLA R. J., Paturel G., FouquéP., 1991, Third Reference Catalogue of have revealed (Chung et al. 2007). The morphology of the tail, as Bright Galaxies. Volume I: Explanations and references. Volume well as the asymmetry and the HI deficiency observed in other II: Data for galaxies between 0h and 12h. Volume III: Data for galaxies in the region suggest that RP is most likely the primary galaxies between 12h and 24h mechanism responsible for the tail. A high velocity collision was Dressler A., 1980, ApJ, 236, 351 also found to be a possible cause of the tail, although the compan- Duc P.-A., Bournaud F., 2008, ApJ, 673, 787 ion remains unidentified. Faridani S., Flöer L., Kerp J., Westmeier T., 2014, A&A, 563, A99 Giovanelli R., Haynes M. P., 1985, ApJ, 292, 404 Giovanelli R. et al., 2005, AJ, 130, 2598 Gunn J. E., Gott J. R. I., 1972, ApJ, 176, 1 ACKNOWLEDGMENTS Haynes M. P., Giovanelli R., Chincarini G. L., 1984, ARA&A, 22, We thank all the team of SKA South Africa for allowing us to ob- 445 tain scientific data during the commissioning phase of KAT-7. The Haynes M. P. et al., 2011, AJ, 170 work of CC is based upon research supported by the South African Henning P. A., Sancisi R., McNamara B. R., 1993, A&A, 268, Research Chairs Initiative (SARChI) of the Department of Science 536 and Technology (DST), the Square Kilometre Array South Africa Hess K. M., Jarrett T. H., Carignan C., Passmoor S. S., Goedhart (SKA SA) and the National Research Foundation (NRF). The re- S., 2015, MNRAS, 452, 1617 search of AS & KH have been supported by SARChI, SKA SA fel- Hess K. M., Pisano D. J., Wilcots E. M., Chengalur J. N., 2009, lowships. AS was additionally funded by the National Astronomy ApJ, 699, 76 and Space Science Programme (NASSP). Karachentsev I. D., Makarov D. I., Karachentseva V. E., Melnyk O. V., 2011, Astrophys. Bull., 66, 1 Kenney J. D. P., Koopmann R. A., Rubin V. C., Young J. S., 1996, AJ, 111, 152 REFERENCES Kenney J. D. P., van Gorkom J. H., Vollmer B., 2004, AJ, 127, Aaronson M., Huchra J., Mould J., 1980, ApJ, 237, 655 3361 Alighieri S. d. S. et al., 2007, A&A, 474, 851 Kent B. R., 2010, ApJ, 725, 2333 Arnaud M., Pointecouteau E., Pratt G. W., 2005, A&A, 441, 893 Kent B. R. et al., 2007, ApJ, 665, L15 Arrigoni Battaia F. et al., 2012, A&A, 543, A112 Kim S. et al., 2014, ApJS, 215, 22 Auld R. et al., 2006, MNRAS, 371, 1617 Koopmann R. A., Giovanelli R., 2008, ApJ, 85 BahéY. M., McCarthy I. G., Balogh M. L., Font A. S., 2013, Koopmann R. A., Kenney J. D. P., 2004, ApJ, 613, 866 MNRAS, 430, 3017 Kurono Y., Morita K.-I., Kamazaki T., 2009, PASJ, 61, 873 Balogh M. L., Navarro J. F., Morris S. L., 2000, ApJ, 540, 113 Mahajan S., Raychaudhury S., Pimbblet K. A., 2012, MNRAS, Barnes D. G. et al., 2001, MNRAS, 322, 486 427, 1252

c 0000 RAS, MNRAS 000, 000–000 10 Sorgho et al.

Makarov D., Prugniel P., Terekhova N., Courtois H., Vauglin I., 2014, A&A, 570, A13 McMullin J. P., Waters B., Schiebel D., Young W., Golap K., 2007, in ASP Conf. Ser., Vol. 376, Astron. Data Anal. Softw. Syst. XVI, Shaw R. A., Hill F., Bell D. J., eds., p. 127 McNamara B. R., Sancisi R., Henning P. A., Junor W., 1994, AJ, 108, 844 Minchin R. et al., 2007, ApJ, 670, 1056 Neill J. D. et al., 2014, ApJ, 792, 129 Nilson P., 1973, Acta Univ. Ups. Nov. Acta Regiae Soc. Sci. Ups. - Uppsala Astron. Obs. Ann. Oosterloo T., van Gorkom J., 2005, A&A, 437, L19 Porter S. C., Raychaudhury S., 2007, MNRAS, 375, 1409 Sandage A., Binggeli B., Tammann G. A., 1985, AJ, 90, 395 Sault R. J., Teuben P. J., Wright M. C. H., 1995, in ASP Conf. Ser., Vol. 77, Astron. Data Anal. Softw. Syst. IV, Shaw R., Payne H., Hayes J., eds. Schwarz U. J., Wakker B. P., 1991, in ASP Conf. Ser., Vol. 19, Ra- dio Interferom. Theory, Tech. Appl. IAU Colloq. 131, Cornwell T. J., Perley R. A., eds., p. 188 Solanes J. M., Giovanelli R., Haynes M. P., 1996, ApJ, 461, 609 Solanes J. M., Sanchis T., Salvador-Sol´eE., Giovanelli R., Haynes M. P., 2002, AJ, 124, 2440 StanimiroviˇcS., 2002, in ASP Conf. Ser., Vol. 278, Single-Dish Radio Astron. Tech. Appl., StanimiroviˇcS., Altschuler D., Gold- smith P., Salter C., eds., pp. 375–396 StanimiroviˇcS., Staveley-Smith L., Dickey J. M., Sault R. J., Snowden S. L., 1999, MNRAS, 302, 417 Taylor R., Davies J. I., Auld R., Minchin R. F., 2012, MNRAS, 423, 787 Tonnesen S., Bryan G. L., van Gorkom J. H., 2007, ApJ, 671, 1434 Toomre A., Toomre J., 1972, ApJ, 178, 623 Tully R. B., Shaya E. J., 1984, ApJ, 281, 31 Urban O., Werner N., Simionescu A., Allen S. W., B¨ohringer H., 2011, MNRAS, 414, 2101 Vogel S. N., Wright M. C. H., Plambeck R. L., Welch W. J., 1984, ApJ, 283, 655 Vollmer B., Beck R., Kenney J. D. P., van Gorkom J. H., 2004, AJ, 127, 3375 Vollmer B., Cayatte V., Balkowski C., Duschl W. J., 2001, ApJ, 561, 708 Warmels R. H., 1988, A&AS, 72, 19 Yun M. S., Ho P. T., Lo K. Y., 1994, Nature, 372, 530

c 0000 RAS, MNRAS 000, 000–000 Observations of Virgo Southern Filament 11

D i d v W c M d Object R.A Dec. Type 25 sys 50 HI defHI M87 ′ −1 −1 8 (J2000) ( ) (deg) (Mpc) (km s ) (km s ) (10 M⊙) (deg) (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11)

NGC 4424 12 27 11.6 09 25 14 SBa 3.63 62.1 15.2 ± 1.9a 433 58.6 2.0 ± 0.4 0.79 ± 0.09 3.10 NGC 4451 12 28 40.5 09 15 31 Sab 1.48 51.2 28.7 ± 1.0 864 255.8 5.4 ± 0.6 0.14 ± 0.17 3.18 NGC 4470 12 29 37.8 07 49 27 Sa 1.29 44.7 16.2 ± 1.6 2321 135.7 1.6 ± 0.3 0.05 ± 0.14 4.58 UGC 7590 12 28 18.8 08 43 46 Sbc 1.35 76.6 21.9 ± 2.9 1112 177.2 16.5 ± 3.5 −0.50 ± 0.17 3.71 NGC 4411A 12 26 30.0 08 52 18 Sc 2.04 54.4 15.1 ± 1.5 1271 105.2 1.8 ± 0.3 0.44 ± 0.09 3.68 NGC 4411B 12 26 47.2 08 53 04 Sc 2.51 26.7 27.9 ± 0.6 1260 153.9 15.6 ± 1.6 0.21 ± 0.06 3.64 UGC 7557 12 27 11.1 07 15 47 Sm 3.02 21.3 16.8 ± 1.7b 924 245.2 4.1 ± 0.7 0.59 ± 0.08 5.21 NGC 4445 12 28 15.9 09 26 10 Sab 2.63 90.0 19.1 ± 0.9 418 171.8 0.4 ± 0.1 1.36 ± 0.14 3.02 VCC 1142 12 28 55.5 08 49 01 dE 0.27 53.4 20.1 ± 2.0c 1334 52.0 0.4 ± 0.1 −0.33 ± 0.33 3.60 NGC 4416 12 26 46.7 07 55 08 Sc 1.70 24.0 36.6 ± 4.0 1381 229.2 3.3 ± 0.6 0.78 ± 0.09 4.58 NGC 4466 12 29 30.6 07 41 47 Sab 1.32 74.9 27.7 ± 3.3 797 185.9 2.1 ± 0.4 0.42 ± 0.11 4.71 UGC 7596 12 28 33.9 08 38 23 Im 1.66 71.9 16.4 ± 1.6d 595 59.5 0.7 ± 0.1 0.82 ± 0.09 3.79 VCC 0952 12 26 55.7 09 52 56 SABc 0.26 54.6 24.8 ± 4.2 1024 100.3 0.9 ± 0.2 −0.63 ± 0.48 2.68 AGESVC1 293 12 29 59.1 08 26 01 ? 0.57e 41.8e 17.0 ± 1.7e 615 87.3 0.2 ± 0.1 – 3.96 Cloud 7c 12 30 25.8 09 28 01 HI cloud – – 16.8 ± 1.7b 496 74.0 0.6 ± 0.1 – 2.93 M49 Cloud 12 29 54.4 07 57 57 HI cloud – – 16.8 ± 1.7b 476 66.0 0.7 ± 0.1 – 4.43 KW Cloud 12 28 34.4 09 18 33 HI cloud – – 16.8 ± 1.7b 1270 73.2 0.7 ± 0.1 – 3.13

Table 3: (1) Object name. (2) Optical positions. (3) Morphological type from RC3. −2 (4) Optical diameter at surface brightness µB = 25 mag arcsec from the RC3 catalogue (see notes). (5) Inclination defined as cos2 i =(q2 q2)/(1 q2) where q is the − 0 − 0 axial ratio of the galaxy and q0 = 0.2 (Aaronson, Huchra & Mould 1980). For NGC 4411 A&B, the inclination was taken from Hyper- LEDA (6) Distance of the galaxy from Sol02 (see notes). (7) Systemic velocity. (8) HI profile width, corrected for instrumental effects, turbulence effects, redshift, and inclination, except for the HI clouds where the correction for inclination has not been applied. (9) The HI mass in solar units, assuming that the HI is optically thin. (10) HI deficiency parameter defined in section 4.2. (11) Projected angular distance from the elliptical M87 in degrees. Notes: a Cortés, Kenney & Hardy 2008, b Distance of M87 with a 10% error, c HyperLEDA (Makarov et al. 2014), d Haynes et al. 2011, e Taylor et al. 2012.

c 0000 RAS, MNRAS 000, 000–000 12 Sorgho et al.

A APPENDIX:HI MAPS AND VELOCITY PROFILES OF THE DETECTIONS

c 0000 RAS, MNRAS 000, 000–000 Observations of Virgo Southern Filament 13

Below we present, from Figure A1 to A13, the HI column density maps of the detections along with their HI proﬁles. For some of the galaxies, namely NGC 4424, NGC 4445, UGC 7557 and UGC 7596, theHI centre is offset with respect to the optical centre. This offset is southwards, probable sign of environment effects. Most of the objects lying in the south-east and north-west of the ﬁeld (see Fig. 4) were not covered by the WSRT pointings, and therefore lack of resolution. NGC 4466 in particular, was partially covered and presents a northern half more resolved than the rest of the galaxy (Fig. A12).

) 1 − ) s / km ( 2 −

cm 17 (10 HI Average N 2 2 velocity (km/s) []

) 2 ) 2

1 1 2 − − ) ) s s

/ / km km ( ( 2 2

− − 2 cm cm

17 17

(10 (10

HI HI

Average N Average N 2 2 2 2 velocity (km/s) velocity (km/s)

Figure A1. Column density map: the contour levels are 0.5, 1, 2, 3 ... 10 ×1019 atoms cm−2 (left contours are NGC 4445 and right are NGC 4424). HI −1 −1 −1 proﬁles: upper right: NGC 4445 (vsys = 418 km s ), lower left: NGC 4424 (vsys = 433 km s ), lower right: NGC 4424’s tail (vsys = 453 km s ).

c 0000 RAS, MNRAS 000, 000–000 ) )

1 1 − − ) )

s s 1 / / km km ( ( 2 2

− − 1 cm cm

17 17 (10 (10 HI HI

Average N Average N 2 11 1 1 1 1 1 1 1 1 velocity (km/s) velocity (km/s)

Figure A2. Column density map: the contour levels of NGC 4451 (blue contours) are 1, 2, ..., 9 ×1019 atoms cm−2 and those of the KW gas cloud (red 19 −2 −1 −1 contours) are 1, 1.5, 2, 2.5 ×10 atoms cm . HI proﬁles: left: NGC 4451 (vsys = 864 km s ), right: KW cloud (vsys = 1270 km s ).

) ) 1 1 − −

) ) s s / / km km

( 1 (

2 2 − −

cm 1 cm 17 17 (10 (10 HI HI Average N Average N 1 1 1 2 2 velocity (km/s) velocity (km/s)

Figure A3. Column density map: the contour levels of NGC 4470 (green contours) are 0.5, 1, 2, ..., 5 ×1019 atoms cm−2 and those of the M49 gas cloud − (blue contours) are 1, 1.5, 2, 2.5 ×1019 atoms cm 2. The dwarf galaxy UGC 7636 (aka VCC 1249) is denoted by a cross south-east of the cloud. HI proﬁles: −1 −1 left: NGC 4470 (vsys = 2321 km s ), right: M49 cloud (vsys = 476 km s ). 16 Sorgho et al.

) UGC 7590 1 − )

s 5 / km ( 4 2 −

cm 3 17

(10 2 HI

0 Average N

600 800 1000 1200 1400 1600 1800 2000 velocity (km/s) c 0000 RAS, MNRAS 000, 000–000

19 −2 −1 Figure A4. Column density map and HI proﬁle of UGC 7590. Contours levels are 2, 5, 10, 20, 30, ... 60 ×10 atoms cm . vsys = 1112 km s . Observations of Virgo Southern Filament 17

) 5 55 1 − ) s / km ( 5 2 −

cm 17 5 (10 HI

5 Average N 5 5 velocity (km/s)

19 −2 −1 Figure A5. Column density map and HI proﬁle of UGC 7557. Contours: 1, 2, ..., 8 ×10 atoms cm . vsys = 924 km s .

c 0000 RAS, MNRAS 000, 000–000 ) NGC 4411A ) NGC 4411B 1 10 1

− − 15 ) ) s s / /

8 km km ( ( 2 2

− − 10 6 cm cm 17 17 4 (10 (10

HI HI 5

Average N Average N 0

600 800 1000 1200 1400 1600 1800 2000 600 800 1000 1200 1400 1600 1800 2000 velocity (km/s) velocity (km/s)

Figure A6. Column density map: the contour levels are 1, 5, 10, 15, 20 ×1019 atoms cm−2 left contours are NGC 4411B and right are NGC 4411A). HI −1 −1 proﬁles: left: NGC 4411B (vsys = 1271 km s ), NGC 4411A (vsys = 1260 km s ). Observations of Virgo Southern Filament 19

) 1 − ) s / km ( 2

− cm

17 (10

Average N 8 8 velocity (km/s)

19 −2 −1 Figure A7. Column density map and HI proﬁle of NGC 4416. Contours: 0.5, 1.0, 1.5, 2.0 ×10 atoms cm . vsys = 1390 km s .

c 0000 RAS, MNRAS 000, 000–000 20 Sorgho et al.

) 1 −

) s / km (

2 − cm 17 (10 HI

Average N 2 2 velocity (km/s)

19 −2 −1 Figure A8. Column density map and HI proﬁle of Cloud 7c. Contours: 0.5, 1.0, 1.5, 2.0 ×10 atoms cm . vsys = 496 km s .

c 0000 RAS, MNRAS 000, 000–000 Observations of Virgo Southern Filament 21

) 1 − ) s

/ km ( 2

− cm 17

(10 HI

Average N 6 6 velocity (km/s)

19 −2 −1 Figure A9. Column density map and HI proﬁle of VCC 1142. Contours: 0.5, 1.0, 1.5, 2.0 ×10 atoms cm . vsys = 1334 km s .

c 0000 RAS, MNRAS 000, 000–000 22 Sorgho et al.

) 1

− ) s / km ( 2 − cm 17 (10 HI Average N 6 6 velocity (km/s)

19 −2 −1 Figure A10. Column density map and HI proﬁle of VCC 0952. Contours: 0.5, 1.0, 1.5, 2.0 ×10 atoms cm . vsys = 1024 km s .

c 0000 RAS, MNRAS 000, 000–000 Observations of Virgo Southern Filament 23

) 1

− ) s / km ( 2 2 − cm 17 (10 HI

Average N 2 2 velocity (km/s)

19 −2 −1 Figure A11. Column density map and HI proﬁle of UGC 7596. Contours: 0.5, 1.0, 1.5, 2.0 ×10 atoms cm . vsys = 595 km s .

c 0000 RAS, MNRAS 000, 000–000 24 Sorgho et al.

) 1

− 5 ) s / km (

2 − cm 17 5 (10 HI

Average N 5 5 5 velocity (km/s)

19 −2 −1 Figure A12. Column density map and HI proﬁle of NGC 4466. Contours: 1.0, 2.0, 3.0, 4.0 ×10 atoms cm . vsys = 797 km s .

c 0000 RAS, MNRAS 000, 000–000 Observations of Virgo Southern Filament 25

) 2 1 − ) s /

km ( 2 − cm

17 (10 HI

Average N 2 2 velocity (km/s)

19 −2 −1 Figure A13. Column density map and HI proﬁle of AGESVC1 293. Contours: 1.0, 1.5, 2.0, 2.5 ×10 atoms cm . vsys = 595 km s . It is not clear whether the complex located south of the object (upper panel) is associated to the cloud, and further study may be needed to investigate this.

c 0000 RAS, MNRAS 000, 000–000 HI

def

4 ◦ dM87 ( )

HI def

6 log M ∗ (M ⊙)